Update README.md

README.md

@@ -57,4 +57,4 @@ Finally the model has mistaken the dock for a boat, which is perhaps understanda
 </p>
 
 ## Conclusions
-With relatively little time (1-2 hours including training) and effort a model was trained that provides decent performance for ships in open water. However small ships, clustered ships, and false positives close to land are an issue. Future work will evaluate the balance of the dataset with regard to these poorly performing cases. Note also that YOLOv5s is the smallest available, and better metrics could be achieved with one of the [larger models](https://github.com/ultralytics/yolov5#pretrained-checkpoints). Experimentation with alternative augmentation strategies could be explored.
+With relatively little time (1-2 hours including training) and effort a model was trained that provides decent performance for ships in open water. However small ships, clustered ships, and false positives close to land are an issue. Future work will evaluate the balance of the dataset with regard to these poorly performing cases. Note also that YOLOv5s is the smallest available, and better metrics could be achieved with one of the [larger models](https://github.com/ultralytics/yolov5#pretrained-checkpoints). Experimentation with alternative augmentation strategies could be explored, to help balance out some of the variations we observe due to: sea conditions (rough or calm), lighting (overcast or bright sunshine), ship density.